Reveal heterogeneous motion states in single nanoparticle trajectory using its own history

Extract latent features of single-particle trajectories with historical experience learning

Single-particle tracking has enabled real-time, in situ quantitative studies of complex systems. However, inferring dynamic state changes from noisy and undersampling trajectories encounters challenges. Here, we introduce a data-driven method for extracting features of subtrajectories with historical experience learning (Deep-SEES), where a single-particle tracking analysis pipeline based on a self-supervised architecture automatically searches for the latent space, allowing effective segmentation of the underlying states from noisy trajectories without prior knowledge on the particle dynamics. We validated our method on a variety of noisy simulated and experimental data. Our results showed that the method can faithfully capture both stable states and their dynamic switch. In highly random systems, our method outperformed commonly used unsupervised methods in inferring motion states, which is important for understanding nanoparticles interacting with living cell membranes, active enzymes, and liquid-liquid phase separation. Self-generating latent features of trajectories could potentially improve the understanding, estimation, and prediction of many complex systems.

Uncover Single Nanoparticle Dynamics on Live Cell Membrane with Data-Driven Historical Experience Analysis

Understanding the spatiotemporal dynamics of particles in a complex biological environment is crucial for the study of related biological processes. To analyze the complicated trajectories recorded from single-particle tracking (SPT), we have proposed a method named SEES based on historical experience vector analysis, which allows both the global patterns and local state continuities of a trajectory to emerge by themselves as color segments without predefined models. This method implements a data-driven strategy and thus uncovers the hidden information with less prior knowledge or subjective bias. Here, we demonstrate its efficiency by comparing its performance with the Hidden Markov model (HMM), one of the most widely used methods in time series processing. The results demonstrated that the SEES operator was more sensitive in identifying rare events and could utilize multivariable observations in the dynamic processes to uncover more details. We applied the method to analyze the dynamics of nanoparticles interacting with live cells expressing programmed death ligand 1 (PD-L1) on the membrane. The results showed that the SEES operator can successfully pinpoint the transmembrane rare events, visualize the on-membrane "Brownian searching" motion, and evaluate different dynamics among multiple trajectories. Furthermore, we found that the PD-L1 expression level on the cell membrane affected the rotation behavior of the nanoparticle as well as the cellular uptake efficiency. These findings enabled by SEES could potentially help the rational design of highly efficient nanocargoes.